Boundary Conditions
 

Dear All,

We have run some diffuser passage simulations with different boundary conditions, all other parameters being the same, just to compare effects of boundary conds.

We used 2 setups:

• Mass flow @ inlet & opening outlet (Ave static outlet pressure = 0)
• Opening @inlet (Ave stat inlet press = 0), mass flow at outlet

RMS residuals fell to ~ 1e-04, not good convergence so we also monitor mass flow average pressure drop through the passage. This settles down to fluctuations of 0.5% full scale, fine for our purposes.

The problem is when we compare results with these two combinations we get almost 30% difference in pressure drop through the passage. I expected negligible difference.

ANSYS were asked if they could explain but we didn't really get to the bottom of it. Maybe we didn't ask properly.

Could anyone offer a reason for this?

Many thanks in anticipation.

Ianto.

P.S. The second combination above was found to give better convergence on an earlier run which is why we looked at these different set-ups to check if there was a difference.

in one case your are "blowing" through your geometry and in the other one you are "sucking", aren´t you?! maybe this could explain the difference.

Thanks for your reply Adam.

Physically yes, but I'm not sure what the difference is numerically. I thought since the same mass is moving through the passage at the same velocity in each case, pressure drop should be the same but it seems I'm wrong. Symptomatic of my only partially understanding what the solver is doing I guess.

Any further suggestions would be much appreciated!

Kind regards,

Ianto

if you are sucking through a diffusor the boundary layer normaly should be attached, if you are blowing it can easily detach. so compare your flow field and not only the pressure change.

Adam,

Sorry for the delay I didn't realise you'd replied a second time, there was no email notification, I was just browsing old posts.

That explanation makes perfect sense it's seems very obvious with hindsight.

Thanks again!

Ianto